Python：類別覆蓋最小採樣個數

import numpy as np
import scipy.io as scio
import pandas as pd
import matplotlib.pyplot as plt
from sklearn import datasets
from sklearn.datasets import fetch_mldata
from sklearn.datasets import fetch_20newsgroups_vectorized


def Cover_RandomSampling(y):
    n = len(y)
    labels = np.unique(y)
    num_labels = len(labels)
    Unobserved = [x for x in range(n)]
    Selected_labels = []
    for i in range(n):
        temp = np.random.choice(Unobserved,replace=False)
        Selected_labels.append(y[temp])
        Unobserved.remove(temp)
        ObservedLabels,ObservedCount = np.unique(Selected_labels,return_counts=True)
        if len(ObservedLabels) == num_labels:
            print(ObservedLabels,ObservedCount)
            break
    return len(Selected_labels)

if __name__ == "__main__":
    # data = np.array(pd.read_csv(r'E:\dataset\clusterData\spiral_1.csv', header=None))
    # data = np.array(pd.read_csv(r'E:\dataset\clusterData\bolbs_hard.csv', header=None))
    # data = np.array(pd.read_csv(r'E:\dataset\clusterData\two_circles.csv', header=None))
    # data = np.array(pd.read_csv(r'E:\dataset\clusterData\aggregation.csv', header=None))
    # data,y = datasets.make_moons(n_samples=1000,shuffle=True,noise=0.1,random_state=101)
    # plt.scatter(data[:,0],data[:,1],c=y,marker='o')
    # plt.show()
    # y = np.vstack(y)
    # X = np.hstack((data,y))
    # X = pd.DataFrame(X)
    # X.to_csv(r'E:\dataset\clusterData\two_moons.csv',header=None,index=None)
    # data = np.array(pd.read_csv(r'E:\dataset\clusterData\sonar.csv', header=None))
    # data = np.array(pd.read_csv(r'E:\dataset\clusterData\proker_label.csv', header=None))
    # y = data[:, -1]

    # mnist = fetch_mldata('MNIST original')
    # y = mnist['target']
    # y = np.loadtxt(r'E:\dataset\clusterData\proker_label.csv',delimiter=',')
    # twenty = fetch_20newsgroups_vectorized(subset='all')
    # y = twenty.target
    #--------------COIL20---------------------#
    # path = r'E:\dataset\clusterData\label.mat'
    # path1 = r'E:\dataset\clusterData\fea.mat '
    # dataA = scio.loadmat(path)
    # dataB = scio.loadmat(path1)
    # X = dataB['fea']
    # y = dataA['label']
    # y = np.hstack(y)
    #--------------------------------------#
    # data = np.array(pd.read_csv(r'E:\dataset\clusterData\flame.csv', header=None))
    # X = data[:, :-1]
    # y = data[:, -1]
    #--------------------------------------#
    # data = np.array(pd.read_csv(r'E:\dataset\clusterData\COIL20_PCA1.csv', header=None))
    # X = data[:, :-1]
    # y = data[:, -1]
    #--------------------------------------#
    data = np.array(pd.read_csv(r'E:\dataset\clusterData\letterABC.csv', header=None))
    X = data[:, :-1]
    y = data[:, -1]

    #################上面是數據##########################distPercent = 20

    iterCount = []
    for i in range(1000):
        count = Cover_RandomSampling(y)
        iterCount.append(count)
    print(np.mean(iterCount))
    print(np.std(iterCount,ddof=1))